One place for every public dataset you'd otherwise scrape
BLS, Eurostat, World Bank, FRED, and 90+ more sources — already ingested, already queryable. Free to explore, Pro when you need private work and agent-scale headroom.
Free
Query every public dataset on Subsets — no scrapers required.
- 90+ public sources: BLS, Eurostat, FRED, World Bank…
- Publish public datasets, charts & articles
- 10 private articles, 50 private charts
- 100 API requests, 50 SQL queries / month
- 10-second query timeout
Pro
Most popularFor analysts and agents that lean hard on public data.
- Everything in Free, plus:
- Unlimited private datasets, charts & articles
- 50× the API and SQL quota for agent workloads
- Priority connector maintenance
- 60-second query timeout
- Pay-as-you-go overage at metered rates
- Priority support
Metered rates
Same rates for everyone. Free accounts prepay through their monthly allowance; Pro accounts get included allowances and pay these rates only on overage.
Data API requests
GET calls to the data endpoint — cached chart and dataset reads.
SQL compute
Each query executed against the DuckDB engine on your datasets.
Bandwidth
Raw data downloads — full parquet and CSV exports.
Frequently asked
How does pay-as-you-go work?
On Pro you get generous monthly allowances. Once you exceed them, usage is billed at the metered rates above. Free accounts stop serving requests when the included quota is reached.
What counts as a data API request?
Any authenticated GET to the public data endpoint that returns rows — including calls from agents, scripts, embeds, and the Subsets frontend itself.
Is my data public by default?
Yes. Subsets is built for agent-native grounding, so datasets, charts, and articles are public unless you mark them private. Free accounts can keep up to 10 articles and 50 charts private.
Can I download my data?
Always. Every public dataset is downloadable as parquet or CSV. Downloads count toward your bandwidth allowance.